===========================================
Description of Oliverarray Annotations
===========================================

Expression data presented here represents at least two hybridized 
slides per species, with RNA from male and female whole flies 
labelled with either Cy3 or Cy5. 

Analysis pipeline:
-----------------

Arrays are scanned and the feature reporst are generated using 
NimbleScan.  There reports are loaded into Bioconductor. 
Variation by channel is minimized using loess normalization 
across the entire slide using the limma package. No background 
subtraction is performed. Between-slide quantile normalization by 
platform is also performed.

Channel specific background intensity is calculated by taking the 
mean intensity values for probes targeting negative control 
elements - probes targeting an Arabidopsis gene AF048702, for 
which no transcript was added to the sample, for each respective 
channel. 

To determine if there was signal detected for each probe, and 
hence a putative transcript present, median probeset intensity 
was compared to background. Percentile ranks of the median 
intensities of all probesets was calculated, and probes with a 
percentile rank greater than background in either channel, in 
each hybridization, were considered above background.  Background 
values were in the 20-25th percentile. The maximum percentile 
rank in any hybridization for each probe is reported in the 
annotation file.

Mapping probes to assembly:
--------------------------

Arrays were designed using sequence versions available in April, 
2005.  Probe sequences were re-mapped to current (CAF1) sequence 
using standalone BLAT v.25x1.  A "MapScore" was calculated for 
each mapping as: (matches - mismatches - gap bases).  Since these 
are 50-mers, the maximum score is 50.  Probes with minimum 
MapScore of 45 were kept.

There may be multiple hits in the assembly per probe.  All hits 
equal to the maximum MapScore are kept.  Probes with > 10 hits 
were removed.


File names:
------------------------------------------------------
The following files contain results for each species:

ana.caf1.gff	D. ananassae
moj.caf1.gff	D. mojavensis
pse.caf1.gff	D. pseudoobscura
pserec.caf1.gff	D. pseudoobscura (reconciled assembly)
sim.caf1.gff	D. simulans
vir.caf1.gff	D. virilis
yak.caf1.gff	D. yakuba
yakrec.caf1.gff	D. yakuba (reconciled assembly)


Field descriptions: 


<seqname>
    The name of the reference sequence (chromosome or scaffold)
<source>
    The source of this annoptation: "Oliverarray"
<feature>
    The feature type name: "mRNA"
<start>, <end>
    Coordinates of best match for this probe
<score>
    Maximum percentile rank of intensity for this probe 
(max=100).  Higher numbers are stronger signals.
<strand>
    '+', '-' 
<frame>
    Not used... "."
[attribute]
    Probe ID given, along with score for the probe re-mapping 
(range 45..50)


===========================================
Design overview for multiple species arrays:
===========================================

Single-slide expression arrays - 380K spot slides from Nimblegen.

Each slide has probes for ~19K gene predictions, with 10 probes 
per gene in duplicate (19K x 10 x 2 = 380K) Probes are not 3' 
biased - they are randomly distributed along the length of the 
gene.

Sources of gene predictions:

EIS: Annotations from Eisen lab (v.1.0, Feb. 2005). Provided by 
Venky Iyer.
GID: ab-initio gene predictions on raw sequence. All gene 
predictions not overlapping with Eisen annotations are kept.
GBK: Individual sequences deposited in GenBank, including ESTs. 
Probes were also designed based on reverse complements for ESTs.
FBA: Annotation from flybase (D.pseudoobscura only - release 
1.03)

Total gene predictions for each type:

		sim	yak	ana	moj	vir	pse
	EIS	13561	12675	12265	9817	11316	12395
	GID	1919	5870	8248	7540	5582	2549
	GBK	843	1273	4	0	92	150
	FBA	0	0	0	0	0	651
	total	16323	19818	20517	17357	16990	15745

Each sequence screened for uniqueness, and probes selected at 
N=10 density (density is lower for some smaller genes).

=================================================================
Updated May. 24, 2006
Dave Sturgill, Yu Zhang, Michael Parisi, and Brian Oliver at the 
Laboratory of Cellular and Developmental Biology, NIDDK, NIH. 
These results are unpublished. Please contact us at 
davidsturgill@niddk.nih.gov if you are interested in using these 
data.